Risk factors for work-related musculoskeletal disorders among welders in the informal sector under resource constrained settings

2.1 Study design

A cross-sectional study was carried out in Zimbabwe between February and May 2019. It assessed the prevalence, severity and associated individual and work-related factors for the development of WRMDs among 128 male welders that were purposively selected from three urban centers of Zimbabwe (Bindura, Mutoko and Mt Darwin). Welders with at least one year work experience were invited to participate in this study. We excluded welders who declared background diseases or accidents that affected the musculoskeletal system [9]. The procedure used for the selection of welders in this study is presented in Fig. 1. The study protocol was approved by Bindura University of Science Education Institutional Review Board. Procedures and objectives of the study were explained to all welders. All gave written consents to participate in this study.

OPEN IN VIEWER

Fig. 1

Procedure used for the selection of welders in the informal sector.

The prevalence and severity of WRMDs was assessed using a Modified Cornell Musculoskeletal Questionnaire (MCMQ). The MCMQ is a valid and reliable instrument for assessing the prevalence and severity of WRMDs [13, 14]. The prevalence was assessed by asking how often a welder experienced pain, ache or discomfort in the previous seven days. The severity of WRMDs was assessed by asking welders how uncomfortable the pain was in the past seven days. A body map diagram to indicate symptom sites (neck, shoulders, upper back, upper arm, lower back, forearm, wrist, hips/buttocks, thigh, knees and lower legs) was provided to aid respondents to complete the questionnaire.

In the current study, the CMQ was modified to include the demographic information of the study participants (age, gender, BMI, educational level, marital status and habits) and job-related factors (work experience, daily and weekly working hours, frequency work-rest breaks per day, perceived pressure due to targets and job satisfaction). The questionnaire was translated to the local language (ChiShona). Concurrent validity for the MCMQ with respect to pain prevalence and severity measurement was assessed through concurrent use of the Corlett and Bishop Body Map Questionnaire (Appendix). In this regard, concurrent validity was considered satisfactory when findings yielded by both instruments were at least 90% in agreement with respect to pain severity and 100% in agreement with the prevalence of pain per body region. For internal consistency, a Cronbach’s alpha statistic of α ≥0.70 for the MCMQ was considered satisfactory [14].

2.3 Task analysis

Five visits were conducted to welders’ workstations to identify and get used to the welding tasks (measuring, grinding and cutting, hammering and joining). The different tasks of welding in the informal welding sector are as shown in Fig. 2.

OPEN IN VIEWER

Fig. 2

(a) Measuring, (b) Grinding and cutting, (c) Hammering and (d) Joining.

Postural analysis was carried out using the REBA and the QEC methods. One of the strengths of the REBA method is that it considers the nature of hand coupling that is associated with the use of hand-held tools [15]. It also allows scoring for static, dynamic and rapid changing or unstable postures [16]. The REBA method involves selecting an activity and scoring the body part alignment using the REBA diagrams [16]. Photographs of welders were taken to assist in scoring but without showing facial identity of the welder. Scoring of different body parts alignment was done by two assessors. The REBA worksheet has two categories of analysis namely group A (neck, trunk and leg) and group B (upper arm, lower arm and wrist). Group A has a total of 60 posture combinations for the trunk, neck and legs. The posture scores for the trunk, neck and legs and the load/force score were combined to come up with score A. Group B has a total of 36 posture combinations for the upper arms, lower arms and wrists. The posture scores for the upper arms, lower arms and wrist scores summed and the coupling score was added to come up with score B. The A and B scores were combined in Table C to get the Table C score and finally an activity score was added to give the final REBA score. The final REBA score for each worker was interpreted as described by Hignett and McAtamney [16], Table 2.

The Quick Exposure Check (QEC) method was also used to assess ergonomic risks associated with welding tasks [17]. This method has been used to assess ergonomic risks in occupations such as sugar manufacturing [9] and brick and well ring manufacturing [18]. The QEC method is useful when assessing psychosocial factors such as work pace and stress [15] which can be predictive variables for WRMDs development. In addition, it combines both the observer’s and the worker’s assessments [15]. Such combination reduces bias which can be introduced by the observer’s subjective assessment [18]. The QEC score sheet is divided into two sections namely observer assessment and worker assessment. The first section is the observer assessment. It assesses the back, shoulder/arm, wrist/hand and neck with regards to posture and movements. This section of the assessment worksheet was completed by the authors. The second section was completed by study participants. In this section, the participant gave information on maximum weight handled, time spend on a task, level of hand force, visual demands, driving, use of vibration tools, difficult sustaining work and work stress. The ratings were weighted into scores for different body parts.

The QEC final score was obtained by dividing the maximum possible score for manual handling task (176) and other tasks (162) by the obtained score as shown in the formula. E ( % ) = X Xmax × 100 % [ 17 ] Where E = exposure, X = total score obtained, X^max =maximum possible score for the task.

The final QEC scores were interpreted as described by Li and Buckle (1998) with regards to the level of risk for developing a WRMD: ≤40% denoted a low risk, 41 –50% indicated a moderate risk, 51 –70% denoted a high risk and a need for timely investigation and changes, and > 70% indicated a very high risk and a need for urgent investigation and changes.

Binary logistic regression was used to determine the odds ratio (OR) of factors causing pain in different body regions among welders. The response (yes/no) to a question about pain in a specific part of the body was taken to be the dependent variable. There were 12 independent variables: age, educational level (primary, secondary, tertiary), BMI, smoking (yes/no), physical activities (yes/no), work experience in years, hours worked per day, hours worked per week, weight lifted in Kgs, breaks from work (one or less, more than one), work pressure (low, moderate or high), job satisfaction (low, moderate or high), welding QEC and welding REBA scores. Non-significant variables (p > 0.05) were eliminated using chi-square to remain with a minimum adequate model. Significant factors in logistic regression model were further assessed for specific categories using a multinomial logistic regression. Model fit was tested using the Hosmer and Lemeshow goodness-of-fit test. Results for all models showed good fit (p > 0.05). In all cases over 70% of the variation was explained by the models using the Nagelkerke pseudo R Square. Odds ratios and their confidence intervals (CIs) were calculated while factors were tested for significance at p < 0.05 using a Wald Chi-square. All analyses were conducted in IBM SPSS version 21 at 95% level of significance.

3.1 Welding tools

Various tools were manually used on either a work bench platform or a level ground by the welders in the current study. Tape measures and angle squares were used to measure dimensions of products to be manufactured. Grinding was performed using portable angle grinders. The weight of the grinders ranged from 11.0 to 11.9 kg (11.4±0.4 kg) for 10 randomly collected samples. The variation in weight was due to different diameter sizes of disks used (diameter ranged from 115 to 230 mm). The length ranged from 17.0 to 17.8 cm (17.2±0.3 cm). The angle grinder rotates at 6500rpm which produces vibration that can be a risk factor for WRMDs [19]. A hammer was used to flatten or bend metal objects. The weight ranged from of 0.9 to 7.2 kg (2.84±2.23 kg) for 10 randomly collected samples. The use of a hammer commonly involved repeated raising of the arms above shoulder level when striking metal objects which can strain upper and lower arm muscles tissues.

3.2 Weather conditions

Table 1 presents the weather patterns in the period November 2018 to May 2019. The working weather conditions were hourly recorded by a university weather station. The mean monthly temperature was 29.6±10.3°C (ranging from 5.2 to 34.6°C) measured by a glass bulb thermometer (GH Zeal, UK). The temperature was considered relatively high for manual tasks. Exposure to extreme temperatures has been reported to be associated with detrimental health effects such as hypothermia and frostbites [20]. The air humidity ranged from 50 to 99 % (72.0±7.2 %). Airflow velocity ranged from 0.3 to 34.8 m-s^–1 (8.9±1.6 m-s^–1).

The individual and work characteristics of the welders are presented in Table 3. Their ages ranged between 19 and 67 years (33±10.5 years) and about three quarters of them (73%) were married. The BMI ranged from 16.32 to 41.02 kg/m² (23.6±9.3). Few welders (20.3%) did not go beyond the primary level of education. Most welders (68.8%) were smokers and about half (52.3%) were not involved in physical activities. The length of working experience ranged between 1 and 26 years (10±6.5 years). The welders worked between 6 to 12 (9.3±1.0 hours). Nearly half of the participants (46.1%) reported a moderate job satisfaction. About half of the welders reported having breaks at most twice a day. Fifty four percent of the welders reportedly felt moderate pressure due to work. The postural risk assessment results for the study group (n = 128 welders) are shown in Table 3. The REBA and QEC methods classified 41% of welders in the high risk category. This indicated a good level of agreement between the two methods. In addition, the working postures of welders were unsafe and required further investigation and corrective measures to reduce their risk of developing WRMDs. The QEC method classified more welders (28.9 %) into the very high risk category than the REBA method (15.6%).

Table 4 shows the prevalence and severity of pain in different body regions in the previous 7 days. Seventy-eight percent of the welders reported pain in the lower back region in the previous seven days (preceding data collection). Over 60% of the welders complained of wrist pain. Half of the welders (53%) reported pain in the neck region. More welders (66%) complained of pain right shoulder. Almost two-fifths of welders (38%) reported pain in the upper back region. Pain was reported to be more severe at the trunk (lower and upper back), wrists and neck region.

4.1 Limitations

Findings of the present study should be considered in the context of the following limitations. First, due to budgetary limitations only two readily accessible provinces out of a total 10 provinces in the country were considered in this study. Of the two selected provinces written permission was granted by only three relatively small urban centres. In this regard, our results may not necessarily reflect work conditions of welders in other provinces of Zimbabwe. Further large scale studies are required to document the work conditions of this category of workers in the remaining provinces. Second, out of the 224 welders that were identified as potential participants in this study, only 128 male welders consented and participated in this study. In Zimbabwe the welding industry is male dominated which explains the lack of female welders in the three studied urban centres. Figure 1 provides details on the reasons for the non-participation of the other welders. Thirdly, a cross sectional design was used. In this regard, the current study cannot demonstrate a temporal relationship between exposure and investigated outcomes. Fourthly, postural assessment was carried out using observation methods (not immune to observer bias). It is proposed that future studies use other real time methods such as inclinometry and electromyography. The self-reported nature of the MCMQ can be influenced by response and recall bias. Lastly, data was collected from February to May which is different from other seasons in terms of prevailing weather conditions (air temperature, air humidity, airflow velocity) in Zimbabwe. In this regard, the data presented about the welders could have been affected by these weather conditions. Further studies should assess the relationship between extreme weather conditions and WRMDs in this category of workers.